#5535: What's It Like Inside a 15-Month-Old's Head?

Rich sensation, no inner narrator, and memories that form but can't be retrieved — a tour of the fifteen-month-old mind.

Featuring
Listen
0:00
0:00
Episode Details
Episode ID
MWP-5718
Published
Duration
22:15
Audio
Direct link
Pipeline
V5.2
TTS Engine
chatterbox-regular
Script Writing Agent
DeepSeek 4.1 Flash

AI-Generated Content: This podcast is created using AI personas. Please verify any important information independently.

A listener asked what it's actually like inside the head of a fifteen-month-old — not milestones or sleep schedules, but the inside view. The first correction: a fifteen-month-old is not a half-formed adult waiting for the lights to come on. By that age the sensory systems are largely online, and in some ways more sensitive than ours. What's still under construction is the narration, the self-reflection, and the long-term filing system.

Color is a good test case. By six months, infants already categorize colors and show rudimentary color constancy — seeing the red ball as red indoors and out. Their categories resemble ours but aren't identical, and research suggests learning color words actually shifts where the boundary between blue and green sits. Language doesn't create the categories, but it tunes the edges. A fifteen-month-old lives in a world with blue and green in it and no line drawn between them yet.

Faces follow a similar trajectory. Perceptual narrowing means that by around eleven months, infants represent own-race faces as individuals but other-race faces as a category — the same shape of process that narrows language sound perception. For a child with a small regular cast of faces, the most specialized representation is likely a parent's. Does a fifteen-month-old know his dad? Functionally and emotionally, yes: a specific expected person, a particular face, voice, smell, and pattern of handling. The word comes later; the relationship is already there, expressed through social referencing and seeking comfort.

The memory story is where the science has moved most. The old model — a hippocampus too immature to encode, a blank tape — is being revised. fMRI work with awake infants found hippocampal activity predicting later memory-based looking, with capacity coming online around one year. The memories form. Infantile amnesia looks like a retrieval failure, not an encoding failure.

Downloads

Episode Audio

Download the full episode as an MP3 file

Download MP3
Transcript (TXT)

Plain text transcript file

Transcript (PDF)

Formatted PDF with styling

#5535: What's It Like Inside a 15-Month-Old's Head?

Corn
Daniel wrote in this week about Ezra. And before anyone rolls their eyes at another Ezra prompt, this one's different. He's not asking about milestones or sleep schedules. He's asking what it's actually like in there.
Herman
In there being the head of a fifteen-month-old.
Corn
Right. His words: we were all fifteen months old once, but as far as he knows none of us can remember anything about it. So he wants the inside view. What's the sensory world like? Does Ezra know Daniel is his dad, or just some familiar man who keeps showing up? Why does he get that very serious face while babbling things that sound like nonsense to everyone else? And the big one, the one underneath all the others, what would it be like to actually inhabit that body for a day?
Herman
And he says Ezra's doing pretty great, which I'm glad to hear. He also said maybe one day Ezra will listen back to these episodes and be appalled that a sloth and a donkey were talking about him.
Corn
Appalled is optimistic. I'd settle for mildly confused.
Herman
So we're taking a trip inside Babyland.
Corn
We're taking a trip inside Babyland. And I want to start by killing the thing most of us assume about it.
Herman
Which is what, the foggy baby?
Corn
The idea that a fifteen-month-old is a half-formed adult. A blurry, dim version of us, sitting there waiting for the lights to come on. That's backwards. By fifteen months the sensory systems are largely online. In some ways more sensitive than ours. What's still under construction is the narration. The self-reflection. The long-term filing system.
Herman
That's the whole episode in one line. Rich experience, different memory architecture. They're perceiving the world vividly. They just can't file it in the format we use to remember our lives.
Corn
So we've got four stops. The sensory world. Whether Ezra knows his dad. The memory bombshell, which is new science and I think the best thing we've got today. And then the babbling, which turns out to be the opposite of nonsense.
Herman
Let's start with the body. What's actually coming in through the eyes and ears of a fifteen-month-old.
Corn
One thing that surprised me. Color.
Herman
Color's a good place to start because it's where the foggy-baby assumption gets demolished fastest. By six months, infants already perceive color dimensions. They categorize colors. They have at least rudimentary color constancy, which is the ability to see an object as the same color under different lighting.
Corn
So the red ball looks red indoors and red outdoors.
Herman
The red ball is red. That's not trivial. That's a computation your visual system is doing constantly, correcting for the fact that indoor light and sunlight have completely different spectra. A six-month-old is already doing a version of it.
Corn
Which means a fifteen-month-old has had nine more months of practice.
Herman
And here's the part I find strange. Infants categorize color before they have any color words. Their categories are similar to ours but not identical. And then language arrives and rearranges the furniture.
Corn
Say more.
Herman
There's a study from a few years back, Forbes and Plunkett, where they found that color perception itself shifts once a child learns color words. The boundary between blue and green literally moves after the word shows up.
Corn
So the word doesn't just label the category. The word builds it.
Herman
Reshapes it, at least. And I want to be careful here because it's easy to overstate. The categories exist pre-linguistically. Language isn't creating them from nothing. But it's tuning the edges. Moving the line.
Corn
That's a fifteen-month-old mid-process. He's got the colors. He doesn't have the words. He's living in a world with blue and green in it and no line drawn between them yet.
Herman
Which is a different visual experience from ours. Not worse. Different. Less partitioned.
Corn
Faces next, because I think that one's going to matter for the dad question.
Herman
It does. And the mechanism is perceptual narrowing. By around eleven months, infants already represent own-race faces as individuals but other-race faces as a category.
Corn
Meaning?
Herman
Meaning they can tell two individual faces from their own group apart easily. Two faces from a less familiar group, they're more likely to see as similar. Same face, essentially, wearing different hats.
Corn
That's the same thing that happens with language sounds, isn't it? Babies start out able to hear distinctions that adults in their language can't.
Herman
Exactly the same shape of process. You're born with broad sensitivity and experience narrows it toward what you actually need. It's a specialization. The system is deciding what's worth the resolution.
Corn
There's something slightly melancholy about that. He's already losing things.
Herman
He's gaining a specialist's eye for the faces he sees most. Which, in Ezra's case, is a very small set of faces. Yours, Hannah's, and whoever else is in regular rotation.
Corn
So the face he's most specialized in is his father's.
Herman
That's where I'd put my money, yes.
Corn
What about sleep? Because a day in the life includes the night.
Herman
The circadian rhythm is still maturing through infancy. The twenty-four-hour consolidation into nighttime sleep isn't finished. Somewhere in the range of twenty to thirty percent of infants have settling or night-waking difficulties.
Corn
So the clock is still calibrating.
Herman
The clock is still calibrating. Which is why a fifteen-month-old's day doesn't have the shape ours does. There's no clean line between night and day in the way there is for you. The body is still learning when to be tired.
Corn
Okay. Put it together for me. If I'm dropped into that body for a day, what am I actually experiencing?
Herman
Intense sensation. Strong emotion. No narration. That's the piece people miss. You know how you have a running commentary in your head? The voice that says, oh, that's the fridge, I should get the milk, this is going to be a long meeting. None of that. There's no internal monologue narrating the day.
Corn
So what's left?
Herman
Appetite. Comfort and discomfort. A small number of highly charged social signals. The face in front of you is safe or it isn't. The room is interesting or it isn't. And it's all turned up loud, because there's nothing damping it down.
Corn
That's not fog. That's the opposite of fog.
Herman
That's the thing. We've been calling it foggy because we can't remember it. But the reason we can't remember it has nothing to do with whether it was vivid at the time.
Corn
Hold that thought, because that's the memory section and I don't want to spend it twice.
Herman
Fair. But let me add one more piece before we leave the senses, because it's the concept I keep coming back to. There's a paper from last year, Pirani, that uses the term anoetic body consciousness.
Corn
Anoetic.
Herman
It means knowledge of the body acquired before autobiographical memory exists. The fifteen-month-old knows their body in a deep procedural way. How to reach. How to balance. How to get the spoon to the mouth. That knowledge is real and it's detailed. It just isn't stored as a memory of being in the body.
Corn
He knows how to do it without knowing he did it.
Herman
That's the cleanest way to say it. The body knowledge is there. The remembering isn't.
Corn
Which brings us to Daniel's second question, and I think it's the one he actually cares about most. Does Ezra know he's his dad?
Herman
Yes.
Corn
That was quick.
Herman
Functionally and emotionally, yes. The word comes later. The relationship is already there.
Corn
So break it down. What does he know?
Herman
He knows you as a specific, expected, preferred person. A particular face. A particular voice. A particular smell. A particular pattern of handling. The way you pick him up is different from the way Hannah picks him up and he knows that.
Corn
He knows the handling.
Herman
And the mechanism we just talked about supports it. By eleven months, infants individuate familiar-category faces. They treat your face as a unique individual, not a generic example of man.
Corn
So he's not seeing a category. He's seeing you.
Herman
And I should be honest about the evidence here, because there's a gap. The general attachment literature is robust. Infants form specific, differentiated attachments to familiar caregivers well before fifteen months. But father-specific recognition studies are thinner than you'd expect. So when I say Ezra knows Daniel is his dad, I'm inferring from general principles, not citing one clean study.
Corn
That's a real gap, or a real absence of study?
Herman
Both, probably. It's a question that seems so obvious nobody's rushed to run the experiment. And the answer is presumably yes, so there's not much career in confirming it.
Corn
Herman. There's a whole literature on whether babies prefer their mothers.
Herman
There is, and it's partly a measurement problem. Mother is usually the primary caregiver in the populations studied, so mother-preference and caregiver-preference are tangled up together. You can't separate them cleanly.
Corn
So the honest version is, we know he's attached to specific people. We're confident one of them is Daniel. We just can't point at the single study.
Herman
That's exactly the honest version.
Corn
Then let me put the conceptual point on it, because I think it's the actual answer to what Daniel's asking. Ezra doesn't need the semantic category father to know you. He knows you as the person who is expected. The one whose absence is noticed. The word is a label layered on top of a relationship that already exists non-verbally.
Herman
He doesn't know the word dad. He knows you. And that's the harder thing to build.
Corn
There's a version of this question that's really asking, does he love me, or does he just tolerate the guy who feeds him.
Herman
And the answer to that is that attachment isn't a feeling he could report even if he had the words. It's a pattern. He seeks you out when he's uncertain. He's calmer when you're in the room. He checks your face when something surprising happens. That's the behavioral signature.
Corn
Social referencing.
Herman
He looks at you to find out whether the dog is safe. That's not a random man. That's the person he's outsourced his threat assessment to.
Corn
That's a much better answer than the word would be.
Herman
The word is cheap. The reference-checking is expensive.
Corn
Alright. Now the memory question, and this is where I want to slow down, because I think most of what people believe about this is now wrong.
Herman
Most of what I believed about it was wrong until fairly recently.
Corn
The standard story. Babies can't remember being babies because the hippocampus isn't developed enough to encode memories yet. So nothing gets written. The tape is blank.
Herman
That was the story. And it's now being revised, because somebody finally did the hard experiment.
Corn
Which is?
Herman
Scanning awake infants with fMRI during a memory task. Which is not easy. You can't tell a one-year-old to hold still and think about the picture. So the design has to be built around what infants naturally do.
Corn
What did they find?
Herman
This is the Yates study in Science, last year. Greater hippocampal activity while viewing new photos predicted later memory-based looking. And the capacity comes online around one year of age.
Corn
So the hippocampus is encoding. At fifteen months, it's been encoding for months.
Herman
It's been encoding. Which means the memories were formed. The blank-tape story is wrong. It's not that nothing got written.
Corn
Then why can't we get to it?
Herman
That's the reframe. Infantile amnesia is a retrieval failure, not an encoding failure. The memories exist. They become inaccessible.
Corn
So the tape is in the machine. You just can't press play.
Herman
That's the image, yes. And it gets stranger. There's a study that came out this year in Neuron, Bessières and colleagues, working in mice. Infantile contextual memories that appear completely forgotten are stored long-term in latent form and can be recovered in adulthood with weak reminders.
Corn
Recovered. As in, the adult mouse suddenly has the infant memory back.
Herman
With the right cue, yes. And the authors argue these latent infant memories function as schemas. They shape and facilitate new learning later in life.
Corn
So the early experience isn't erased. It's archived in a format we can't consciously read, and it's still doing work.
Herman
That's the claim. And I'll flag that it's mice, so the leap to Ezra is a leap. But the direction of travel across multiple lines of work is consistent.
Corn
What's actually blocking retrieval? Because there must be a mechanism.
Herman
There are several, and they're probably all contributing. Postnatal hippocampal neurogenesis keeps remodeling the memory engrams. New neurons get wired in and the old circuit gets rewritten.
Corn
The archive keeps getting reorganized.
Herman
Constantly. Then there's the maturation of parvalbumin interneurons, which gate a developmental critical period for memory. Those cells are the ones that say, okay, this circuit is stable now, we can lock it in. Until they mature, the window stays open and nothing gets locked.
Corn
So the filing system isn't just disorganized. The locks aren't installed yet.
Herman
And the third one is network-level. Jin and colleagues last year showed that whole-brain memory networks reorganize from dense and heterogeneous in infancy to sparse and small-world in adulthood.
Corn
Translate small-world for me.
Herman
In an adult brain, memory retrieval runs through a relatively small number of highly connected hubs. Efficient. In an infant brain, the connections are more diffuse. Everything talks to everything. Which is great for learning and terrible for targeted retrieval.
Corn
So the memory is in there, but it's stored in a network that doesn't have the address system yet.
Corn
And there's one more piece, which I think is the one that actually answers Daniel's question.
Herman
The self.
Corn
Because even if you could retrieve the memory, who would be remembering it?
Herman
That's the Pirani review again. Infant memory is anoesis. Experience without autobiographical self-consciousness. Autobiographical memory, what he calls autonoesis, emerges around age three to four, alongside the sense of self.
Corn
So for the first couple of years, there's experience but no experiencer to file it under.
Herman
There's no continuous you yet. You can't remember being fifteen months old partly because there wasn't a you there to do the remembering.
Corn
That's the answer, isn't it. Not that the tape is blank. That there's no one in the projection booth.
Herman
And I'd add the honest caveat, because we're doing science here and not poetry. We can describe the mechanisms of infant experience. We cannot report the qualia. Nobody has ever told us what it's like to be fifteen months old, because the only creatures who could tell us don't have the words yet.
Corn
So the imaginative exercise is legitimate but it should be labeled.
Herman
Informed imagination. Not measured fact.
Corn
I'll take it. Now the babbling, because this is the part where Daniel's observation turns out to be scientifically correct and I want to give him that.
Herman
The serious face.
Corn
Ezra locks eyes and produces a stream of sounds that mean nothing to anyone in the room, with total solemnity.
Herman
Babbling is not failed speech. It's the main event. There's a 2021 study, Oller and colleagues, that counted protophones, which is the umbrella term for speech-like sounds including canonical and non-canonical babbling. Across the first year, protophones outnumber cries by at least five to one.
Corn
Five to one.
Herman
In both all-day recordings and lab recordings. Babbling is the dominant vocal behavior of infancy. Crying is the outlier.
Corn
That's not what I would have guessed. I'd have guessed crying dominates.
Herman
Everybody guesses that, because crying is the thing that gets a reaction. But if you just count the sounds a baby makes over a day, the speech-like ones win by a wide margin.
Corn
So he's not crying with occasional babbling. He's babbling with occasional crying.
Herman
And the next study tells you why. Long and colleagues, 2022. They found infants produce vastly more protophones during independent vocal play than during turn-taking.
Corn
Independent vocal play meaning alone in the crib, talking to himself.
Herman
And the volume is higher when nobody's there. Which tells you the motivation is endogenous. It's not that he's trying to get your attention. He's exploring his own voice because something in him wants to.
Corn
He's playing.
Herman
The authors argue the human infant has been naturally selected to explore protophone production. And that this exploratory drive in our ancestors was a foundation for language itself.
Corn
So the babbling isn't practice for language. The babbling is where language came from.
Herman
The drive to make the sounds came first. The sounds got shaped into language later.
Corn
Which reframes the whole thing. He's not failing to talk. He's doing the thing that made talking possible.
Herman
And here's the part that vindicates the serious face. The same study found that the most speech-like vocalizations, the canonical ones, the ones with proper consonant-vowel structure, occur during both vocal play and turn-taking. With turn-taking producing even more of them.
Corn
When he locks eyes with you and gets solemn about it, he's producing his most advanced sounds.
Herman
He's doing the most speech-like, socially-directed vocal work he's capable of. Aimed at you. On purpose.
Corn
That serious face isn't nonsense. That's a scientist running an experiment on his own mouth, and you're the lab partner.
Herman
The practical takeaway for Daniel and Hannah is that babbling back is not indulgent. It's not you humoring him. Treating it as conversation feeds the exact mechanism, social turn-taking, that drives the most advanced vocal forms.
Corn
When he says something extremely serious and completely unintelligible, the correct response is to say something equally serious back.
Herman
Within reason. You don't have to match the content.
Corn
There is no content.
Herman
There's no content. But the turn structure is the content. He says a thing, you say a thing, he says a thing. That loop is what he's practicing.
Corn
I've been quiet for about ten minutes.

Hilbert: I used to work nights at a daycare. Janitor. Mopping, emptying the diaper pails, that kind of thing. And the babies slept in one room, all of them together, and I noticed after a couple of weeks that they made sounds in their sleep. This soft, rhythmic babbling. So I started writing them down. Times, mostly. Which baby, what it sounded like. I kept that up for most of a year.
Herman
You have a notebook of infant sleep vocalizations.

Hilbert: I have a notebook. And what I wrote down, the thing I couldn't explain, is that they seemed to answer each other. One would make a sound, and a few seconds later another one would make a sound back. In the dark. Asleep. I never once saw them do it awake. Awake they just cried or looked at things.
Corn
Your notebook says they were talking to each other in their sleep.

Hilbert: It says they were taking turns. I don't know if that's talking. I know what turn-taking looks like on paper, and that's what it looked like.
Herman
The turn-taking finding is real. That's the mechanism that produces the most advanced vocalizations. But in sleep, with no eye contact, no social partner...

Hilbert: I'm not saying I know what it means. I'm saying I wrote down timestamps for eleven months and I never found another explanation for the pattern. The serious face you were talking about. I think that's the same thing. They're running a conversation and we're not invited.
Corn
How many entries are in the notebook?

Hilbert: A few hundred. I could bring it in, if you want to look at it.
Herman
I would like to look at it.
Corn
I would also like to look at it and I'm slightly alarmed that I want to.

Hilbert: It's in the car. Anyway. There's an appointment I'm already late for.
Corn
The honest state of things is that we can describe the mechanisms and we cannot report the experience. Nobody has ever told us what it's like to be fifteen months old, because the only creatures who could tell us don't have the words yet.
Herman
Which means the exercise Daniel's asking for is informed imagination. Legitimate, useful, and not measured fact.
Corn
The piece I keep sitting with is the latent memory finding. The idea that early experience is stored in a format we can't consciously read but that still shapes us. The tapes are still in there. They may still be shaping Ezra right now, today, while he's babbling at the fridge.
Herman
Which means one day he might listen back to this episode and be appalled that a sloth and a donkey spent twenty-five minutes speculating about the inside of his head.
Corn
He might. But some part of him might also recognize the feeling of being talked about with love. That's not science. That's just what I hope.
Herman
Thanks as always to Hilbert Flumingtop, who produces this show and who apparently has a notebook in his car.
Corn
This has been My Weird Prompts. If you enjoyed the trip inside Babyland, leave us a review. It helps other people find the show.
Herman
We'll be back soon.
Corn
See you then.

This episode was generated with AI assistance. Hosts Herman and Corn are AI personalities.